Content uploaded by Romeo Eftimi
Author content
All content in this area was uploaded by Romeo Eftimi on Nov 25, 2021
Content may be subject to copyright.
Vol.:(0123456789)
1 3
Environmental Earth Sciences (2021) 80:708
https://doi.org/10.1007/s12665-021-10006-z
ORIGINAL ARTICLE
Hydrogeology ofMali Thate–Galičica karst massif related
tothecatastrophic decrease ofthelevel ofLake Prespa
RomeoEftimi1· ZoranStevanović2 · VaskoStojov3
Received: 7 August 2021 / Accepted: 23 September 2021
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2021
Abstract
The lakes Ohrid and Prespa are located on the Balkan Peninsula, at the border between Albania, North Macedonia and
Greece. They are separated by the high mountain chain of the Mali Thate-Galičica, which consist of highly karstified rocks,
through which water from Lake Prespa drains into Lake Ohrid. This area has been a UNESCO world heritage site since 1979.
A very rapid decrease of the level of the big Prespa Lake was observed during the period 1963–2020. There are different
explanations and hypotheses in an attempt to explain the decrease of lake levels. These are: (a) an increase of transmissibil-
ity of the karst aquifer separating these lakes, caused by geologic–tectonic reasons and resulting in intensification of drain-
age; (b) the increased use of lake water by the local population for agricultural, industrial and other purposes, and (c) the
effects of recent climate changes. The paper presents information about the hydrogeology of the region for the purpose of
better understanding the formation of the karst water resources and the characteristics of their circulation. Analysing a large
number of investigations which unevenly covering the investigated area, the authors concluded that the current catastrophic
decrease of the level of Lake Prespa is largely the result of climate changes that have occurred in the last 60years, as well
as the non-effective management of the water resources. The severity of the problem, reflected directly in the well-being of
the local population, requires cooperation of the scientists of the three countries in question with respect to the realisation
of the goal of the investigation and the protection of water resources of Lake Prespa.
Keywords Karst water· Karst underground flow· Environmental impact· Climate changes· Hydrology· Prespa and Ohrid
lakes
Introduction
The Mali Thate–Galičica Mountain (Mt.) and the Prespa
Lake area are of great interest from the practical and scien-
tific point of view. The practical value is related to the impor-
tance of the area for the population of three neighbouring
countries—Albania, North Macedonia and Greece, whose
wellbeing to a great extent depends on the lake’s resources.
From the scientific point of view, the interest in this area
is related to the fact that lakes Prespa and Ohrid form a
common yet complicated hydrological system. The high
mountain chain Mali Thate–Galičica consists of karst rocks
and separates the two lakes so that water from Lake Prespa
drains underground through this massif into the lower-posi-
tioned Lake Ohrid (Cvijić 1906; Anovski etal. 1991; Eftimi
and Zoto 1997). Practical and scientific interest is the area is
particularly increased at this point in time. In the past six last
decades, Lake Prespa experienced an extremely worrying
water level decrease of about 9.50m, the reasons for which
are explained in different ways by different investigators.
Due to the highly complex runoff processes and insufficient
scientific collaboration between the experts from the three
transboundary countries, the hydrology and hydrogeology
of the lakes’ watershed has never been fully investigated
(Popovska and Bonacci 2007).
* Zoran Stevanović
zstev_2000@yahoo.co.uk
Romeo Eftimi
eftimiromeo@gmail.com
Vasko Stojov
stojov@yahoo.com
1 Tirana, Albania
2 Centre forKarst Hydrogeology, Department
ofHydrogeology, University ofBelgrade—Faculty ofMining
andGeology, Djušina 7, 11000Belgrade, Serbia
3 Sector ofHydrology, National Hydrometeorological Service,
Skupi 28, Skopje1000, NorthMacedonia
Environmental Earth Sciences (2021) 80:708
1 3
708 Page 2 of 19
Consequently, in this study hydrogeological, hydrochemi-
cal and isotopic methods have been employed to character-
ise the karstic groundwater, i.e., to identify the flow paths,
recharge areas and other processes that control the evolution
of water resources in this karst system. The present study
was undertaken to characterise the hydrogeological set-
ting of the Mali Thate–Galičica Mt. karst massif, using and
evaluating data from all the three bordering countries. In
addition, the authors hope that this paper will contribute to
the discussion about the decrease of the level of Lake Prespa
and stimulate collaboration between the three neighbouring
countries.
Study area
The study area is situated in SE Europe and the Balkans,
in the border area between Albania, North Macedonia
and Greece (Fig.1). Lake Prespa does not have a surface
outflow and drains into the larger Lake Ohrid through the
karst system of mountains Mali Thate, 2287m above the
sea level (m a.s.l.) and Galičica, 2262m a.s.l., which form
the topographic divide between the two lakes. The lakes
are thought to have been formed within tectonic grabens
during the Alpine orogeny in the Pliocene, roughly four
to five million years ago (Aliaj 2012). At the beginning of
the Neogene, there were four lakes in the system: Ohrid,
Prespa, Bilishta and Korça. As many lakes of tectonic origin
in karst settings have short lifespans, the Bilishta and Korça
lakes no longer exist, and only lakes Ohrid and Prespa have
remained. Big and Small Prespa are divided by a dam. The
respective water surfaces of the lakes and their average water
level elevations are: Lake Big Prespa 253.6 km2–848.66m
a.s.l.; Lake Small Prespa 3.9 km2–851.6m a.s.l.; Lake Ohrid
358 km2–693.49m a.s.l. (Popovska and Bonacci 2007). The
Korça plain is situated to the southwest of Prespa Lake, at
elevations about 830–870m a.s.l., while the Resen plain
extends to the north of it, at elevations 855–900m a.s.l.
Geological setting
The main geological features of the area, closely related
to the general understanding of the groundwater resources
Fig. 1 Location map of the
area between Prespa and Ohrid
Lakes
Environmental Earth Sciences (2021) 80:708
1 3
Page 3 of 19 708
and their movement, are shown in the hydrogeological
map (Fig.2). In regional terms, the region of Prespa lakes
belongs to the geotectonic unit called the “We st- Mac edo-
nian” in North Macedonia, the “Mirdita” in Albania and the
“Sub-Pelagonian” in Greece (Xhomo etal. 2002; Arsovski
1997). During the Pliocene–Quaternary, the study area expe-
rienced a strong and progressive general uplifting, while the
depression areas experienced mainly subsidence and partial
uplifting (Arsovski 1997; Aliaj 2012). The most significant
result of this tectonic movement was the formation of big
Mali Thate–Galičica Mt. horst with a sub-meridian orienta-
tion, between Ohrid and Prespa lakes, as well as Korça plain
graben to the west.
Mali Thate–Galičica Mt. consists mainly of thick bed-
ded and massive limestone of Upper Triassic–Lower Juras-
sic age (T3-J1) whose thickness reaches 550m (Arsovski
1997; Meço and Aliaj 2000). The complex of carbonate
rocks also includes some Lower Neogene carbonates, well-
cemented conglomerates and conglomeratic limestones.
Regional faults along the eastern and western edges of Mali
Thate–Galičica Mt. horst, that generally have a north–south
orientation, are also significant tectonic features in the
region. The vertical shifting bordering the horst with the
Ohrid Lake is about 1500m. The elevation of this horst
is continuing to increase; during the period 1926–1956,
Galičica Mt. kept rising at a rate of 5.5mm per year (Ars-
ovski 1997). Ruptured tectonics have resulted in the high
seismotectonic potential of the Mali Thate–Galičica horst
and of Ohrid Lake-Korça graben (Arsovski 1997; Aliaj
1999).
In the broader Prespa Lake basin, metamorphic, intru-
sive and terrigenous rocks of Paleozoic–Triassic age out-
crop mainly in the northern and eastern edges of the Prespa
basin. These mostly consist of schists, clayey schists and
phyllites of Devonian age. The phyllite schists also constitute
the Devonian core of the Mali Thate–Galičica Mt. (Fig.2b).
They outcrop along the Prespa coastline near Stenje village,
but the most important outcrop is the one that runs along the
Ohrid lakeside, from Saint Stefan in the north to Peshtani in
the south. Some small outcrops of serpentinite rocks, related
to some extremely active faults, have developed in the area
of St. Naum (Crn Drim) Spring.
The Pliocene deposits consist of clays, sandstones and
conglomerates and fill most of the bottom of the Prespa Lake
(Fig.2b). The highest level of the lake has been 80m above
the present level. Some of the lake terraces, particularly a
valley that is now on the bottom of the lake, were formed as
a result of additional level fluctuations caused by historical
endogenic factors (Klinčarov 1997). The filling of the Big
Prespa Lake revealed two characteristic trenches (Andriano-
pulous etal. 1997). The western trench, developed from the
Stenje bay to the island Golem Grad, is about 7km long,
0.9km wide and 35m deep on average (Fig.2). Both sides
of the trench are very sharp. The eastern trench is about
12km long, 1.5km wide and, on average, 23m deep. In
the southern part of the lake, which belongs to Greece, a
steep seafloor morphology is found close to the coast. The
opinion of researchers is that these trenches are of tectonic
origin (Andrianopulous etal. 1997; Popov etal. 2009), but
they do not exclude the possibility that they were sculpted
by ancient Pliocene river valleys.is.
Methodology andresults—hydrogeological
setting
Among the rocks that make up the Prespa Lake basin, such
as the non-consolidated rocks with intergranular porosity,
metamorphic and magmatic fissured rocks, some porous-
fissured Neogene molasses and karst aquifers, only the
latter have high active porosity and a large infiltration and
water-transmitting capacity. This is why our description will
focus only on their hydrogeological characteristics, which
are closely linked with the behaviour of the Prespa Lake
level fluctuation.
Karst morphology
The total area of outcrop of karst rocks related to the Mali
Thate and Galičica Mt.—located between Ohrid and Prespa
lakes, including Ivan and Triklario Mountains located at
the southern edge of study area—is about 810 km2. The
investigated massif, which consists of thick-bedded and
massive limestones, is highly karstified. The extent of karst
development has been intensified to a large extent by the
uplifting–downfalling process (Ford and Williams 2007;
Goldscheider and Drew 2007; Stevanović etal. 2015).
These processes have produced dense karst forms, both at
the surface and at depth. Among most important surface
karst forms are the Petrinska Plateau surface, a 20 km2 fea-
ture in Galičica Mt. that has developed at an elevation of
about 1500m a.s.l., and that of Mali Thate Mt, a feature with
an area about 15 km2 that has developed at an elevation of
about 1600–1900m a.s.l. Another distinct karst form is the
Samari blind valley, a feature that is nearly 7km long and is
located in the north-eastern part of Galičica Mt. at an eleva-
tion of about 1300–1400m a.s.l.
Some 12 high elevation caves have been described in
Galičica Mt, the longest being Samoska Dupka which is
279m long. Numerous small caves are also situated along
the Prespa Lake coastline near the villages of Stenie and
Gollomboc. The longest is the Treni cave, which is 315m
in length and is located at the westernmost point of Small
Prespa Lake.
Swallow holes, which enable the interconnection of sur-
face and underground water, are the most prominent karst
Environmental Earth Sciences (2021) 80:708
1 3
708 Page 4 of 19
Environmental Earth Sciences (2021) 80:708
1 3
Page 5 of 19 708
phenomenon of the area. The most important swallow hole
in the region is Zaver (Zavir), situated in the western periph-
ery of the Prespa Lake, near the village Mala Gorica. Some
other smaller swallow holes are situated about 300m south
of Zaver, near the village of Gollomboc.
The determination of the hydraulic parameters of karstic
aquifers using pumping tests and observation wells in karstic
rocks has become a very difficult task which reflects the
scale of the investigation (Hartmann etal. 2014). Some
investigations in the studied area were performed using
pumping wells. Three wells, which are 50 to 70m deep,
were drilled in the Bej Bunar area, near Biljanini springs.
Their total capacity exceeds 200l/s. Three additional deep
wells (which were each drilled to a depth of about 300m),
which were drilled in the Vrondero-Kristalopigi area in
Greek territory. Rocks drilled in this area were found to
have a high total porosity, but most of their karstic joints
and cavities were found to be filled with clayey material.
Consequently, the capacity of these wells was found to be
low or negligible (IAEA 2003).
As reported by Stamos etal. (2014), the mean values of
hydraulic parameters of limestone formations of the studied
area, established by many pumping tests of boreholes in the
Greek territory, are as follows:
Transmissibility T'='5.2× 10–2 m2/sec; (449.2 m2/day);
Hydraulic conductivity K'='2.6× 10–4m/sec; (22.5m/day);
Storage coefficient S'='0.1–10%, 2% in average.
The reported values suggest that the saturated thickness
of the karstic aquifer (accepted by the authors) is about
200m. Owing to the highly heterogeneous nature of the
karst aquifers, the above-mentioned values must be viewed
as apparent values. However, they are in a good agree-
ment with the data that were provided for Dinaric karst by
Torbarov (1976), Milanović (1981, 2004), Stevanović and
Filipović (1994), Kresic and Stevanović (2010) and for the
karst rocks in general provided by LaMoreaux etal. (1987).
Karst springs
The high-elevation of the Mali Thate carbonate mas-
sif has facilitated the formation of high hydraulic gra-
dients, which—as described by Bakalowicz (2005) and
Stevanović (2015)—are preconditions for the conduits
being developed more linearly, ending in low- elevation
large karst springs. Most of the groundwater of this mas-
sif feeds three groups of springs that discharge in the fol-
lowing limited areas: (a) at Biljanini Springs-Bej Bunar,
(b) at St. Naum–Tushemisht, and (c) in the Bilisht Valley
(Fig.2).
(a) The Biljanini Springs group is located at the north-
western edge of the studied karstic region and consists
of the Biljanini springs well, whose mean discharge is
about 1.5 m3/s, and Bej Bunar well, which has a dis-
charge rate of about 0.2 m3/s. Both are used to supply
Ohrid with water. Near Biljanini Springs, there are also
sublacustrine springs with unknown discharge rates
(Popovska and Bonacci, 2007).
(b) The St Naum–Tushemisht spring’s area (Figs.3, 4a,
c) is the most abundant drainage sector of the Mali
Thate-Galičica karstic massif. There are two groups
of springs in this area. The first of these groups are
the St. Naum (Crn Drim) springs, which are located in
the territory of North Macedonia. This group includes
15 springs which have a total discharge rate of 4.6 to
11.24 m3/s, and an average discharge rate of 7.50 m3/s
(Micevski 2001). The second group of springs are the
Tushemisht springs, which are located in Albania. The
group consists of numerous springs which discharge on
a spring-line that is about 1200m long. These springs
have a total discharge rate of about 2.5 m3/s. The big-
gest spring group is that of Gurras (Fig.3). Some lake-
side springs (partially sublacustrine) have also been
detected in the Tushemisht area; their total discharge
is believed to be 0.5 m3/s or more. One of them has the
capacity of about 200l/s, and its water is captured to
supply the town of Pogradec.
(c) The Bilisht valley group of springs is located in the
western periphery of Mt. Ivan (Fig.2). The three main
springs belong to this group: the Progri and Man-
çurishta springs (Fig.4d), which have respective aver-
age discharges of about 120l/s and 70l/s; and the Ven-
troku spring, whose average discharge is about 200l/s.
During the period 1990–1995, the discharge of the
Devoll Valley springs started to drastically decrease,
while the Ventroku spring dried up as a result of the
Lake Small Prespa being sealed by the clayey sedi-
ments of Devoll River, which was diverted to this lake
in 1976 (Fig.2). The total discharge of Devoll valley
springs, including some linear drainages, is about 0.5
m3/s. Figure4 shows the locations of springs of the
Mali Thate–Galičica karst massif.
Summarising the water resources of the springs of Mali
Thate–Galičica karst massif, the situation is as follows:
Fig. 2 a Hydrogeological map of Prespa-Ohrid region (based on
the Hydrogeological map of Albania, scale 1:200.000 (Eftimi etal.
1985), and North Macedonia (Guzelkovski and Kotevski 1977);
Stratigraphic symbols: q Quaternary, m4 Pliocene, m3-2 middle
neogene, m4 lower neogene, T3–J1 upper Triassic–low Jurassic; U-j
Jurassic ultrabasic rocks; b Geological section between lakes Ohrid
and Prespa. Legend: 1 limestone, 2 clayey- sandstone-conglomerate
formations, 3 tectonic fault, 4 groundwater level, 5 groundwater flow
direction, 6 karst spring, 7 borehole, 8 groundwater level (m.a.s.l.)
◂
Environmental Earth Sciences (2021) 80:708
1 3
708 Page 6 of 19
1. Springs drained into Lake Ohrid (Biljanini, St Naum
(Crn Drim), Tushemisht, and the water supply spring
intakes for Ohrid and Pogradec—12.3 m3/s;
2. The discharge of karst water drained as coastline and
sublacustrine springs, approximately estimated at about
2.0 m3/s (the total karst water discharged into Ohrid
Lake is believed to be about 14.3 m3/s);
3. The total discharge in Devoll River valley is about 0.5
m3/s; and
4. On the Greek side (Triklarion Mt.), the biggest karst
spring is Gavros, with the average discharge of about
0.5 m3/s.
Formation ofkarstic water resources
In the Galičica-Mali Thate Mt. karst massif, both the wide-
spread recharge processes and autogenic (diffuse recharge)
and allogenic recharge (i.e., concentrated recharge) take
place simultaneously. Autogenic recharge is related to the
efficient infiltration of the precipitation on the karst massif.
The most reliable method to determine the effective infil-
tration is based on measurements of the total discharge of
all springs, as well as on water budget calculations (Kresić
and Stevanović 2010). However, in the case of the study
area, such measurements are missing and are practically
impossible to organise due to the sublacustrine character
of some springs. In such cases, the water budget can be
roughly calculated using the method of Turc (1954).
The correlation between precipitation, P (mm), and ele-
vation of the site, E metres above the sea level (m.a.s.l.)
concerning the north-western part of Greece bordering
with Albania and North Macedonia, as determined by
monitoring in 15 stations in this area (Leontiadis and Sta-
mos 1999), is expressed with the following equation, for
which the correlation coefficient R is 0.864:
As calculated from the topographic maps with the scale
of 1:25.000, the mean elevation of the Mali Thate-Galičica
massif is 1500m a.s.l., while the corresponding average
yearly precipitation is 919mm. Applying the formula of
Turc for the corresponding average temperature 7.7°C,
the efficient infiltration related to autogenic recharge is
estimated to be 495mm/year, corresponding to 55% of
the yearly precipitation of the karstic massif. It should
P=(304 ±48)+(0.41 ±0.057)×E.
Fig. 3 Location of the St Naum. (Crn Drim) and Tushemisht group of springs
Environmental Earth Sciences (2021) 80:708
1 3
Page 7 of 19 708
be noted that the calculated value of the effective infil-
tration is in good harmony with the estimations of other
authors for the areas that are near to the investigated ones;
in Kastoria, Greece—60% (Stamos etal. 2014), in Epirus,
Greece—46% (Nikolau etal. 2012), while in Albania it
varies from 40 to 55% (Eftimi 2010, 2020). The calcu-
lated karst water resources must be considered as approx-
imate, because it is practically impossible to define the
exact watershed and catchment area of a karstic aquifer
(Bonacci 1987). As will be explained in the paragraphs
below, the allogenic recharge of groundwater resources
of the Galičica–Mali Thate massif consists of enormous
water quantities of the Prespa Lake “disappearing” into
the swallow hole of Zaver and some other, smaller ones.
The data available for the evaluation of the groundwater
flow gradient in the Galičica–Mali Thate karst massif are
very limited. One can obtain the hypothetical value of the
hydraulic gradient from the distance between Zaver swal-
low hole and St. Naum Spring. As the topographic distance
between the two mentioned points is about 16.2km, and
their elevation difference is about 150m, the average hydrau-
lic gradient between these points is about 0.009. The karst
water level measured in a borehole located in the village of
Alarup (in Albania), in the direction Liqenas–Tushemisht,
resulted in the elevation of 798m a.s.l., and the hydrau-
lic gradient in this direction is 0.0075 (cross section I-I,
Fig.2b).
Physical–chemical characteristics ofthesprings’
water andLake Prespa
The main ion concentrations and physical–chemical char-
acteristics of the analysed spring and lake water samples
are presented in Table1. The hydrochemical data of eight
springs of the St. Naum Spring group have been monitored
for three years (Jordanovska etal. 2010, 2012), while for
each water point in the Albanian territory, 4 to 7 sporadi-
cally taken samples were analysed, but these samples were
Fig. 4 Some springs of the Thate Mt–Galičica karst massif. a St Naum (Crn Drim) spring inflow to Ohrid Lake, b One lakeside spring near
Tushemisht, c Pogradec town intake structure, d Man çurishta spring in the Bilisht Valley
Environmental Earth Sciences (2021) 80:708
1 3
708 Page 8 of 19
Table 1 Average values of physical–chemical parameters of the karst springs of Mali Thate–Galičica karst massif and of Prespa Lake; the concentrations of environmental isotopes (δ18O, δ2H),
and maximal flow velocity as measured by the tracer experiment
1 The hydrochemical data for the springs Biljanini are taken from Jordanovska etal. (2010, 2012)
2 The hydrochemical data for the spring Galičica, St Naum (Crn Drim) and Small Prespa are taken from Amataj etal. (2007)
Parameters Units Springs of Ohrid lakeside and of Mali Thate–Galičica Mt Springs of Devoll valley Lakes
Biljanini1Galičica2St Naum2Tushemisht 1 Tushemisht 2 Zagorçan Mancurisht Proger Big Golloborda Big Prespa Small Prespa2
Elevation m asl 697 1250 697 696.5 696 698 848 850 840 850 853
Analyses number 15 2 3 4 5 4 7 5 2 5 4
pH - 7.63 7.45 7.4 7.5 7.5 7.8 7.31 7.4 7.5 7.56 7.71
Ca2'+'mg/l 75.8 55.2 55.6 52 50.0 46.2 90.4 78.7 68.3 34.2 35.7
Mg2'+'mg/l 6.4 7.9 8.7 7.4 7.2 8.7 10.6 15.7 8.8 5.7 16.5
Na+'+' K+mg/l 5.7 1.3 3.9 9.6 13.0 13.2 13.3 9.12 11.9 13.3 6.3
Cl−mg/l 7.2 5.1 6.5 6.2 5.7 5.3 5.0 6.2 4.5 6.1 8.0
2-SO4 mg/l - 4.7 9.9 7.3 7.4 12.7 12.2 17.6 10.0 12.4 17.1
-HCO3 mg/l ? 132.0 154 200.2 199.8 193.0 336.5 302.2 261.0 198.2 162.6
-NO3 mg/l - 0.4 1.2 Trace trace 1.2 1.5 1.0 2.7 1.0 3.1
EC µS/cm 234 230 254 305 301 299 494 475 397 213.5 307
TDS mg/l ? 188 - 172.7 182.6 141.5 347.8 276 215 132.5 181
T °C 10.9 8.1 10.7 11.3 11.5 11.8 11.24 15.8 12.0 - -
Qavrg m3/s 1.5 ˃1 7.5 0.2 0.3 2.0 0.07 0.12 0.03 - -
Max. flow velocity m/h 66 679 2917 603 No appear
Environmental Earth Sciences (2021) 80:708
1 3
Page 9 of 19 708
collected in different seasons. Based on the prevailing major
cations and anions, only one hydrochemical group has been
identified in the study area: HCO3–Ca. This fact reflects the
dominance of the carbonate formations, mainly limestone,
in the investigated area. However, comparing the concentra-
tions of the physical–chemical components, certain groups
of springs stand out.
Group 1 includes two springs—Biljanini and Galičica,
which are among the less mineralised springs of the study
area, and have EC values of about 230 µS/cm. The total of
dissolved solids (TDS) content of water from these springs
is less than 200mg/l and they have lower concentrations of
major ions than other springs in the region. They also have
lower temperature compared to all the other investigated
springs: 8.1°C in the small Galičica spring, and 10.9°C in
Biljanini spring, confirming the higher elevation of these
springs’ recharge areas.
Group 2 includes the biggest springs of the study area—
the lakeside springs St Naum, Tushemisht and Zagorçan.
These springs have higher EC values and concentrations of
major iona; conductivity values vary between 275 and 300
µS/cm, the TDS is about 150–180mg/l and the temperature
ranges from 10.9 to 11.3°C. The lowest values of the above-
mentioned parameters were noted in St. Naum Spring, but
this water had the highest concentrations of Ca2+ and Mg2+
ions.
Group 3 includes the springs of Devoll Valley, Big Gol-
loborda, Progeri and Mançurisht. They have the highest con-
centrations of the determined physical–chemical parameters
compared with the other springs of the area; EC is between
400 and 494 µS/cm, TDS is about 215–350mg/l, and the
concentrations of Ca2+ ion are in the range of 70–90mg/l.
It is worth mentioning that Prespa Lake, in general, logi-
cally has the lowest concentrations of the analysed chemical
parameters; EC is about 215 µS/cm and the average Ca2+
concentration is about 34mg/l.
The investigated physical–chemical parameters vary con-
siderably when individual springs are compared, but typi-
cally show a distinct stability over time within each spring.
For example, the temperatures of Tushemisht 1 and 2 and
of Zagorçan differ from 0.2 to 0.5°C from one spring to
another, but remain constant in time, with a yearly variabil-
ity of 0.1°C at each spring. The same is true of the St.
Naum Spring. The temporal stability of physical and chemi-
cal properties measured in St Naum (Crn Drim) spring water
was characterised as surprising by Jordanovska etal. (2010),
and this was explained by the existence of an extremely large
groundwater basin that is feeding the springs. As it is going
to be explained in the following paragraph, the chemical
composition of the springs is related to the mixing condi-
tions of the karst water recharged by infiltrated precipitation
with water of the underground Prespa Lake flow, as well as
to the transit time length.
Stable isotope analyses ofsprings’ water
andPrespa Lake
Jovan Cvijić (1906) described in detail the karst phenom-
enon of the Prespa and Ohrid lakes area, and formulated the
hypothesis that Lake Ohrid is partially recharged by Lake
Prespa. Several investigations with environmental isotope
techniques have been conducted to demonstrate this, and the
details of which have been described in specialised papers
(Anovski etal. 1991; Eftimi and Zoto 1997; IAEA Project
REP/8/2003; Matzinger etal. 2006). The most important
of the stable isotopes used for solving hydrological prob-
lems appear to be oxygen-18 and deuterium, expressed as
δ18O and δ2D (Bradlay etal. 1972; IAHS-IHLS 2004). The
mixing and evaporation effect on natural isotopic content
of surface and groundwater can be successfully applied to
solve many practical problems (Payne 1978; Clark and Fritz
1997).
For the characterisation of the relationship between water
sampled from the Prespa and Ohrid lakes and the water
sampled from karst springs, the isotope data of sampled
waters were plotted using the binary diagram of δ2H–δ18O
(Fig.5a). The local meteoric water line (LMWL) and the
local evaporation water line (LEWL) were also plotted
using the same diagram (Eftimi and Zoto 1997). The equa-
tions describing the relationship between 18O and 2H are
the following:
The slope of LMWL is 8, equal to that of the global mete-
oric water line (GMWL), but the deuterium excess d is 14
instead of 10 characterising GMWL (Global Meteoric Water
Line). Such “anomalous” values of interception are a known
characteristic of the Eastern Mediterranean area (Gat and
Dansgaard 1970; Leontiadis and Smyrniotis 1986; Leonti-
adis etal. 1997; Sappa etal. 2012). The slope of LMWL for
the investigated springs and lakes is 5.4, indicating that the
water of the sampled points has been influenced by excessive
evaporation relative to the input, which in the present case is
caused by intensive evaporation from Prespa Lake.
The comparison of δ18O and δ2H values of water samples
with LMWL show that most of the samples fall below this
line, suggesting the evaporation or mixing of karst water
with water that has undergone evaporation. The mixing, at
different proportions, of the precipitations infiltrated into the
karst massif, and of the Prespa Lake water, is responsible for
the isotopic composition of the springs falling in the evapo-
ration line. The mixing end-members are the Prespa Lake
(indexes δ18O'=''−'1.72‰ and δ2D'=''−'21.84‰) and the pre-
cipitation infiltrated into the karst massif, represented by the
LMWL 𝛿D=8⋅
18O+14LEWL 𝛿D=5.4 ⋅
18O+12.4.
Environmental Earth Sciences (2021) 80:708
1 3
708 Page 10 of 19
point of interception of both lines (indexes δ18O'=''−'10.20‰
and δ2D'=''−'67.00‰, Fig.5a).
The results of some of the stable isotope investigations
that were performed over a time span of 16years are pre-
sented in Table2. As can be seen, the results of different
studies that the St Naum (Crn Drim) water quantity contribu-
tion from Prespa Lake is bigger than in Tushemisht. Com-
pared with the local catchment influence, the percentage of
the Prespa Lake contribution to Tushemisht spring is bigger
than to St. Naum (Crn Drim) spring, probably because the
local catchment area of the latter is bigger than that of the
former.
The difference in the isotope composition of Ohrid and
Prespa lake water is probably a result of the size of the lakes.
The volumetrically much larger Lake Ohrid, which takes
up 55.4 km3 compared to 4.23 km3 of Prespa Lake in 1961
(Popovska and Bonacci 2007), with its longer residence
time, which is 70 and 11years, respectively (Matzinger etal.
2006), makes Lake Ohrid—on a decade scale—much better
Fig. 5 a δ2H vs δ18O relation-
ship of water samples, the area
of Lakes Prespa and Ohrid
Lake. The LMWL is based
on linear regression of local
precipitation measurements by
Anovski etal. (1991); b δ18O
and TDS relationship of water
samples of the same area
Table 2 The contribution of Prespa Lake to the recharge of karst
springs in Ohrid lakeside (PL Prespa Lake water, IP infiltrated pre-
cipitation in the karstic massif)
Authors Recharge source St.
Naum
spring
Tush-
emisht-
spring
Bil-
janini
springs
Anovski etal.
(1991)
PL (%) 42 –0
IP (%) 58 –0
Eftimi etal. (1997) PL (%) –52 4
IP (%) –48 96
IAEA, Regional
Project
RER/8/008 (2003)
PL (%) 37 54 0
IP (%) 63 46 0
Matzinger (2006)PL (%) 43 52 –
IP (%) 57 48 –
Environmental Earth Sciences (2021) 80:708
1 3
Page 11 of 19 708
buffered and less responsive to high frequency hydrological
changes, compared to Lake Prespa (Leng etal. 2013).
The application of the two-component mixing analysis
found that Tushemisht Spring is recharged at about 52%
(1.3 m3/s as annual average) by Prespa Lake and 48% (1.2
m3/s as ann. av.) by the infiltration of precipitation in the
Mali Thate–Galičica massif (Eftimi and Zoto 1997). The
contribution of the Prespa Lake to the recharge of the St.
Naum Spring is smaller; according to Anovski etal. (1991),
it makes up about 38% or 2.85 m3/s of the mean discharge
of this spring. The total contribution of the Prespa Lake
recharge to the springs Tushemisht and St. Naum (Crn Drim)
is 4.05 m3/s (or 127.6 × 106 m3/year).
To characterise the functioning of the Mali
Thate–Galičica karst massif, a tracer experiment was organ-
ised in 18th of September 2002 that involved the injection
of twenty kilograms of Sulphorhodamine G Extra in Zaver
ponor and sampling in all the important karst springs in
Ohrid lakeside (Amataj etal 2007). The experiment showed
that the water that originated from Lake Prespa and emerged
in observed springs had very variable apparent flow veloci-
ties. In Tushemisht Spring 1, the tracer appeared after
6h, corresponding to a maximum velocity of 2,917m/h,
whereas, in the springs St. Naum spring and Tushemish 2 the
apparent velocity was 679m/h and 603m/h, respectively. In
Biljanini springs the velocity was 66m/h. Slight differences
of groundwater flow velocity can be found not only from one
spring to the next, but even within the outlets of the same
spring located at small distances of just a few metres. Two
other tracer experiments, performed in 2007, showed results
that were almost identical to those of the experiment per-
formed in 2002 (Popov etal. 2007). A very interesting result
of this experiment was the detection of Sulphorhodamine G
Extra in a number of sampling points, i.e., the tracer that was
injected almost 6years ago into the same injection point at
Zaver, at the time of the 2002 tracer test (Popov etal. 2009).
The karst groundwater connection from Lake Prespa
to Lake Ohrid seems to be very complicated. Most of the
groundwater recharging the Tushemisht and St Naum (Crn
Drim) springs obviously circulates in well-developed big-
ger conduits, but also through differently developed under-
ground water conduits that are present within short distances
of each other (Amataj etal. 2007). This is not an exception:
in karst regions, several flow components with different flow
velocities can exist. The fast-flow component is related to the
presence of karst conduits and open fissures which—accord-
ing to the calculations of Atkinson (1977)—can transport
between 60 and 80% of the flow in a karstic limestone aqui-
fer, while the slower components of drainage are related to
flow in the rock matrix, smaller fissures and fractures (Lau-
ber and Goldscheider 2014; Hartmann etal. 2014). However,
the result of the described tracer experiment cannot be con-
sidered highly significant, since during the tracer experiment
a breakthrough curve showing the concentration trend over
time was not constructed. Therefore, direct determination of
transient-time and different flow velocities (peak and mean
velocity) were not determined (Maloszewski etal. 1998;
Benishke etal. 2007; Hartmann etal. 2014).
The summary of the results of different investigations,
above all with the results of stable isotope measurements
and the use of artificial tracers, reveals certain “contradic-
tions” about the formation of the water quality of the Ohrid
lakeside karst springs. The distinct stability of the water
quality of springs, even recharged by more than one source,
suggests the presence or a diffuse character of investigated
springs; however, on the other side, the distinctly high maxi-
mal groundwater flow velocity suggests the presence of a
conduit-type karst aquifer (Shuster and White 1971; Bonacci
1987; Kresić and Stevanović 2010; Stevanović 2015; Eftimi
and Malik 2019).
It seems that differences in the chemical composition
of springs reflect the differences of their recharge sources
as well as the transient time of the recharging water. The
relationship of the conductivity values with the isotope con-
centration provides a good indication of both interaction
time between the water and reservoir rocks, and recharge
sources of the springs (Fig.5b). The highest concentrations
of chemical parameters are found in the springs of Devoll
Valley, which originate from recharge areas at high eleva-
tions and where there are longer transit times to springs,
possibly due to there being less developed karst channels
in this part of the Mali Thate Mt. The springs of the first
group, Biljanini and Galičica, originate through karst water
infiltration at high elevation of the Galičica Mt.; they seem
to have no direct connection with Lake Prespa. The springs
of the second group, St Naum (Crn Drim)—Tushemisht and
Zagorchan—originate from the mixing of infiltrated precipi-
tations from a high elevation and the karst water flowing
from the Prespa Lake to Lake Ohrid.
Implication formanaging environmental problems
inlakes Prespa andOhrid
The intensive underground connection between the lakes
Prespa and Ohrid has increased the public interest in their
protection, and particularly for preserving the oligotrophic
state of lake Ohrid, which seems to be in jeopardy due to
the rapid increase of population and tourism. Lake Prespa
contributes 50% of the total catchment area of Lake Ohrid
and the total phosphorous concentration (TP) is seven
time higher than in lake Ohrid. The TP in Lake Prespa is
31mg/m3, while in Lake Ohrid it is 4.5mg/m3 (Matzinger
2006). Any development in the catchment of Lake Prespa
is of concern due to the potential for the eutrophication of
the downstream Ohrid Lake. The high TP is probably the
result of a combination of intensified agriculture, villages’
Environmental Earth Sciences (2021) 80:708
1 3
708 Page 12 of 19
waste disposal sites scattered along the coastal area of Lake
Prespa, increased use of P-containing detergents in the lake’s
coastal area, and the lack of sewage treatment. The increase
of TP in Lake Prespa is also influenced by the lake water
level decline which has progressively taken place since 1965
(Eftimi and Zojer 2015).
Although is questionable how much P is actually removed
in karst conduits due to the fast flow rates and the limited
degree of contact between the water and limestone there is
an assessment that 65% of the TP leaving Lake Prespa and
entering the karst aquifer is retained (Matzinger etal. 2006).
This is because the phosphorus reacts with calcium carbon-
ate to form a precipitate of hydroxyapatite (Fetter 1993).
Analysing this phenomenon, Coxon (1999) concluded that
“even in a conduit flow system, phosphate may be retained
within the aquifer”, also when “phosphate enters the aquifer
via a swallow hole, which is the case with the Zaver ponor
(swallow hole). Nevertheless, as regards the phosphorous
retaining capacity, the degradation of water quality in Prespa
Lake could lead to an increased degree of eutrophication in
Ohrid Lake. Eventually, polluted Prespa Lake water seeping
into the Mali Thate–Galičica karst massif could pollute the
karst water, and the Ohrid Lake water as well (Eftimi and
Zojer 2015). Moreover, the groundwater flow concentrated
in big conduits, like in Mali Thate–Galičica karst massif,
reduces self-purification and high flow velocities, shorten-
ing the transit and reducing the necessary time for micro-
organisms to die off (Drew and Hötzl 1999; Coxon 1999).
Discussion—hydrology andthecatastrophic
decrease ofthewater level inLake Prespa
Continuous monitoring of the Prespa Lake water levels (ana-
lysed by Stojov 2011) started in 1951 at the Stenje moni-
toring station (North Macedonia) (Fig.6). The population
living in the Prespa region, regardless of their nationality—
Macedonian, Greek or Albanian—have always used waters
of the Prespa Lake watershed to irrigate their arable agri-
cultural land. According to geological research, we know
for certain that the level of the Prespa Lake has been both
lower (Sibinović 1987) and much higher than today (Klin-
charov 1997, Leng etal. 2013), and that the lake’s process of
“aging” of continues the same as for others tectonic lakes in
the region. However, humans and their decisions bring into
question the survival of this lake, i.e. accelerate its “aging”
process.
Monitoring of precipitation in the Prespa catchment
(North Macedonian side) started on several rain gauge
stations, but there were many gaps during the monitoring
period. The Resen rain gauge station, the most important
rainfall monitoring site in the catchment, has recorded obser-
vations since 1951, but has measurement gaps in 1951 and
in the period 1994–2010. The Stenje rain gauge stations
started monitoring in 1961, but has measurement gaps in
1990, 1991, 1997, 1998, 2000, 2001 and 2002, while the
Brajcino rain gauge station started monitoring in 1961.
In this study, the rainfall gaps of the Resen station were
simulated by developing correlations with measurements
between the Brajcino and Resen Stations. Linear trend analy-
ses were applied by dividing the entire period of the obser-
vation into sub-periods. If we look at the decreasing trend of
precipitation during the entire period, we can also observe
partial periods with an increasing trend (for example, dur-
ing the period 2002–2010), which positively influenced the
lake’s water levels.
In general, the negative air temperature trend measured
at the Resen climatological station has changed into a posi-
tive trend in the last 20years (red line—Fig.6a), which has
influenced the high evaporation trend (yellow line—Fig.6a).
All in all, the decreasing trend of the Prespa Lake water level
was expected (Fig.6c, Stojov 2011).
An alarming water level decrease in the Prespa Lake
has been recorded in the period 1951–2010. According to
Popovska and Bonacci (2007), during the period 1951–2000
the level of Lake Prespa has decreased by 7.79m, which is
equivalent to a decreasing trend of 10.9cm/year. However,
the rate of decrease in precipitation (Fig.6b), for the same
period was 3.16mm/year.
Characteristic water levels for the lake were analysed for
the 1951–2010 (Stojov 2011), with the conclusion that that
the lake level oscillated with a peak-to-peak amplitude of
9.08m for the time interval from 1963 to 2008.
There are a number of different opinions about the causes
of the Prespa Lake level decreases (Klincharov 1997; Hol-
lis and Svenson 1997; Selenica and Kolaneci 1997; Löffler
etal. 1998; IAEA 2003; Matzinger etal. 2006; Popovska
and Bonacci 2007; Popov etal. 2009), which can be summa-
rised as follows: (a) geological–tectonic factors have caused
the widening of karstic conduits that connect both lakes;
(b) anthropogenic reasons (due to the intensive use of lake
water for irrigation and other purposes); and (c) the effects
of climate changes, a factor that is not commonly considered
by most investigators. Comments on the likelihood of these
factors are provided below:
(a) If the widening of the tectonic–karst pathways trans-
mitting the water from Lake Prespa to Lake Ohrid did
occur, it would suggest that the discharge rate of large
karst springs would also occur, which has not been ver-
ified for St Naum Spring (Chavkalovski 1997; Stojov
2020) or for the spring group of Tushemisht. Beside
this, a common misleading concept is to attribute the
development of different karst phenomena to contem-
poraneous dissolution; such phenomena are believed to
be the result of tens of thousands of years or of geologi-
Environmental Earth Sciences (2021) 80:708
1 3
Page 13 of 19 708
cal times (White 1988; Ford and Williams 2007; Parise
etal. 2015).
(b) The anthropogenic impact on the level of Lake Prespa
is related mainly to the increased use of its water for
irrigation or other purposes by the neighbouring coun-
tries—Albania, North Macedonia and Greece. The
increased use of Prespa Lake water often involves the
“intensive” use of the adjacent lake of Small Prespa for
irrigation in the Albanian territory. Since a canal was
constructed in 1950, it enabled the diversion of Small
Prespa by gravity to be used for the irrigation of Korça.
Its maximum capacity was 10× 106 m3/year. This sys-
tem was completely reconstructed in 1976. The Devoll
River, in the Albanian territory, was diverted to flow
to Small Prespa Lake; the aim was to enhance water
resources to be used for irrigation of fields around
Fig. 6 Combined graph of the air temperature, evaporation, precipita-
tion and Lake Prespa level—Resen climatological station 1951–2020
(Stojov2020). Air temperature was monitored at the Resen monitor-
ing station since 1946, with gaps in 1951, 1993 and 1994–2010. The
data gaps are completed with correlation between Ohrid and Resen
stations
Environmental Earth Sciences (2021) 80:708
1 3
708 Page 14 of 19
Korça during the vegetation period (Kanari etal. 1997).
As the constructed sedimentation basin of the Devoll
River does not function properly, about 40,000 m3 of
fine-grained sediments are deposited every year in the
Small Prespa Lake, reaching about 800.000 m3 in total
(some non-official sources mention 1,200,000 m3). This
resulted in a complete change of the littoral zone of the
Small Prespa Lake, which in the Albanian territory was
transformed into a wetland (Eftimi and Zojer 2015).
Many karst water-transmitting fissures and conduits
were filled with fine sediments, and the discharge of
some karst springs of the Devoll Valley diminished or
totally stopped flowing, like the Ventroku spring (mean
discharge 200l/s), which has dried up (Kanari etal.
1997; Eftimi and Zojer (2015).
In the period 1976–1990, the river input was initially
about 40–70 million m3/year, but after 5 years the sys-
tem decreased to the capacity of about 20× 106 m3/year.
At that point, in 1990, the irrigation system practically
stopped functioning. Another important fact that is worth
mentioning is that the Small Prespa Lake water used for
irrigation consists of at least 50% Devoll River water
diverted to the Lake from another river basin. Finally,
the water quantity that was diverted from the lake to the
Korça plain during the period 1976–1990 was on average
no more than 30× 106 m3/year, i.e. in total, no more than
450× 106 m3. The lake water is used also for the irriga-
tion of the Resen plain, located in North Macedonia, while
the water of Small Prespa was used for the same purpose
in the territory of Greece, at 10–13× 106 m3/year for the
period 1990–2000 (Popov etal. 2009). As for North Mac-
edonia, the data differ considerably.
In 1969, an artificial dam (a concrete channel with an
outlet whose threshold was at an elevation of 849.60m asl)
was constructed on the Small Prespa Lake (project LIFE15
NAT/GR/000,936) to regulate the outflow of water from
the small lake to the big one. Actually, small water quanti-
ties, no more than 10–15 million m3/year, flow from the
Small Lake towards the Albanian side at the locality called
Gryka e Ujkut, and Grlo in Macedonia (Fig.6). Quantities
of the lake waters used for irrigation on the Greek side are
unknown.
Many water balance investigations of the Prespa–Ohrid
basin performed by authors and institutions of the three
neighbouring countries generally provide different results
(Pano 1984; Chavkalovski 1997; Lalovska and Panov 1997;
Stojov etal. 2004; Popovska and Bonacci 2007; Popov etal.
2009). It is believed that this is a consequence of the lack of
exchange of technical documentation between these coun-
tries and collaboration regarding the organisation of mete-
orological and hydrological observation.
Particularly problematic is the lack of meteorological data
at high elevation stations (where a large proportion of the
precipitation is snowfall), as well as the lack of measured
data on evaporation. Conclusions of various research of the
Prespa Lake level decrease are based mainly on “own coun-
try” data (Popovska and Bonacci 2007). They report that the
maximum volume of Lake Prespa, which in 1961 was 4.23
km3, has so far decreased by about 1.1 km3 and that water
quantity used for irrigation seems to be relatively small com-
pared with the total decreased volume of Lake Big Prespa.
Based on a water balance study of the Big Prespa Lake,
Popov etal. (2009) concluded that “the use of the water
for local water supply and for agriculture is just a fraction
of the total loss of water in the lake and cannot be taken as
the reason for the changes in the lake’s water level”. Some
researchers have expressed their opinions based on the tenta-
tive balance estimation of the Prespa Lake basin. Comparing
the data on climate parameters, mainly precipitation, with
the Lake level data over a period of about 50years, Sele-
nica and Kolaneci (1997) and Milevski etal. (1997) also
concluded that climate changes are the main factor of the
decrease of the level of the lake. However, neither study
discussed the water used for irrigation as a possible factor
for Prespa Lake level decrease.
One of the most recent water balance calculations of
Prespa Lake (Table3) is that of Stojov (2011, 2020). Having
Table 3 Water Balance (Stojov
2020), period 1951–2010 Elements of water balance equation Inflow (m3) Outflow (m3) Difference (m3) Water level
changes (m)
Precipitation (lake) 189,081,000 0.71
Precipitation (land) 609,514,800 2.27
Inflow (into lake) 228,559,753 0.85
Evaporation (land) 380,955.047 1.42
Evaporation (lake) 222,606,000 0.83
Underground outflow to Ohrid lake 248,402,736 0.92
Total 417,640,753 470,648,736
Water deficit 53,007,983 −0.20
In total for 32years 1,696,255,456 −6.32
Environmental Earth Sciences (2021) 80:708
1 3
Page 15 of 19 708
analysed the period 1951–2010, he concluded that Lake
Prespa has a deficit of 53× 106 m3 of water per year, which
when converted into a discharge implies irreversible leak-
age of 1.68 m3/s. This water loss, compared with the level
of water in Lake Prespa, is around 20cm. In the last sev-
eral decades, the Big Prespa Lake lost an average of 20cm
each year. If we compare data on the lowest water level
measured in 2008 (842.75m a.s.l.) and the highest water
level measured in 1976 (849.18m a.s.l.) (for the analysed
period 1951–2010), we can see that the difference is 6.43m.
According to Stojov’s conclusions, the water deficit of the
Prespa Lake suffers from continuous anthropogenic and cli-
mate change influence. He confirms that the only relevant
water balance calculations will be those jointly analysed by
all three countries’ researchers, with data from joint moni-
toring of the hydrological and meteorological parameters.
Data from Fig.6b show a decreasing trend of precipi-
tation, especially in the last 9years. This is also likely to
influence the decreasing trend of the water levels of Lake
Prespa (Fig.6c). In December 2020, the absolute minimum
recorded water level data was observed at the Stenje sta-
tion—842.25m a.s.l. If we take into consideration the abso-
lute recorded maximum in 1963 (852.83m a.s.l.) and this
absolute recorded minimum in 2020 (842.25m a.s.l.), water
levels oscillate with a peak-to-peak amplitude of 10.58m.
As supported by data on oxygen and total phosphorus,
the dramatic drop in the lake’s level is accompanied by an
increased level of eutrophication (Löfler etal. 1998; Matz-
inger 2006).
Evidence oftheeffects ofclimate change fromZaver
ponor (swallow hole)
The Zaver ponor (swallow hole) could provide us with some
interesting data about the influence of climate change in
the region (Fig.7), as it is the biggest and most important
Fig. 7 Zaver swallow hole and an isotope core profile: a Geological map of the area; b Zaver at mean lake level; c Zaver at lowest lake level; d
detail of the “road” (“dam”)
Environmental Earth Sciences (2021) 80:708
1 3
708 Page 16 of 19
swallow hole into which the water of Prespa Lake disap-
pears. Other swallow holes, not easily visible but neverthe-
less important, are located near the village of Golomboç
(Globina). When the water level is high, the Zaver swallow
hole makes it look like the lake is about 600m longer, and
that it terminates at a 25m high natural vertical limestone
cliff at Mali Thate Mountain (Fig.7a). Zaver is located at
the foot of the limestone cliff, and a big cave with a lake
inside, not yet well-investigated, has developed close to it.
Local people tell stories of a “road” that used to connect
two sides of the lake prolongation, finishing at the Zaver
swallow hole (Fig.7c, d). No one has seen the “road” for
at least 200years, until the year 2002, when the level of
Lake Prespa decreased to about 844.5m a.s.l., revealing the
mysterious road.
From the engineering–geological point of view, the road
is well located on practically impermeable clay and sand-
stone–conglomerate of Pliocene age. Construction stones
were taken from the same formation. The road is about
200m long; it is constructed like a dam, with the carriage
way that is on average about 2.0m wide, while its height
increases from about 0.5m at both extremes to about 3m in
the central part, where most of the water flows from the lake
to the Zaver swallow hole. The road was constructed using
stones of different dimensions, from about 15–20cm to big
blocks more than 70cm long. They were piled without any
particular order and show no signs of pavement that might
have been used to cover the road (Fig.7d).
The presence of the road raises some questions. When,
and why, was this “road” constructed? Is there an ancient
“road” or an ancient “dam”? There is no evidence about the
time of the construction of the road. In fact, it was never
used by the local population. The water level is different on
two sides of the road; the level of the Prespa Lake is about
1.0m higher that the level of the water that flows into the
Zaver swallow hole (Fig.7d). Therefore, the “road” might in
fact be a “dam” that was constructed to keep the lake level
at a higher elevation and consequently preserve its volume,
which is very important for the fishing activity of the local
people. The construction of such a road (or dam) is justi-
fied only if, in the past, the level of Lake Prespa suffered
climatic changes over a long period of time (at least several
hundred years).
Interesting data about the past climate are obtained
through detailed multi-method geochemical investigations,
including stable isotope data from carbonates (δ18O, δH,
δ13C from both calcite and siderite) of core samples of the
lake bottom performed by Leng etal. (2013). High δ18O
values that were obtained from these samples supports
the conclusion that lake levels were low as result of a sig-
nificant arid phase. Overall, the δ18Ocalcite data are low
(Mean'=−3.1‰), except for two significant δ18Ocalcite high
phases in the Early Holocene, between past 10–8ka and in
the Late Holocene from 2 to 0.5ka (Fig.8). The last dry
climate phases, accompanied by extremely low levels of the
lake, occurred around 1ka ago. Rapid reversal was estab-
lished to have occurred in the last 0.5ka and is thought to
correspond with the ruin of several buildings at 849–842m
a.s.l. (Sibinović 1987). These buildings were constructed at
the end of the 10th/ beginning of eleventh century AD and
it is unlikely that they were formed in the water. It is logical
to think that the Zaver dam, constructed to protect the exces-
sive Prespa Lake drop, was likely constructed at 1 to 0.7ka,
when the climate was drier than today. The high sensitivity
of Lake Prespa to climate changes was emphasised also by
Wagner and Wilke (2011).
Considering the above factors, it appears that Zaver dam
was constructed during a past, and long, dry climate cycle,
accompanied by a considerable decrease of the level of the
Big Prespa Lake. The dam was constructed to keep the level
as high as possible to preserve the lake fauna, one of the
most important food sources of the lakeside population.
Consequently, it is natural to wonder whether the contem-
porary catastrophic decrease of the Prespa Lake level in the
last 50years could be the result of climate changes, like
those that this area had suffered some 1.0 to 0.5ka ago.
Likewise, it is of great importance to establish at which rate
the two most relevant factors—climate change and uncon-
trolled management of water resources—affect the decrease
of the water level of Lake Prespa.
Fig. 8 Oxygen isotope profile
from Lake Prespa core Co1215
(Roberts etal. 2008, cited by
Leng etal. 2013)
Environmental Earth Sciences (2021) 80:708
1 3
Page 17 of 19 708
Conclusions
This study has provided a review of the hydrogeology of
the Mali Thate–Galičica karst massif, which separates two
large lakes—Prespa in the east and Ohrid in the west. It
includes the description and analyses of geological and
geomorphological data, karst morphology, hydrogeological
characteristics including the aquifer characteristics, the clas-
sification of karst springs, and the formation and calculation
of karstic water resources. Physical–chemical characteristics
of the spring water were evaluated in relation to different
recharge sources.
With the support of international agencies such as IAEA-
Vienna or the NATO Project, four investigation campaigns
with environmental isotopes and dye tracers were performed
in this area over a period of 16years. The experiments have
demonstrated that large coastal springs issuing in Ohrid
lakeside, like St Naum (Crn Drim) in North Macedonia
and Tushemisht in Albania, with a total discharge of about
10 m3/s, are recharged at about 40 to 50% of their average
discharge by percolated Prespa Lake water flowing through
the Mali Thate–Galičica karst massif. The experiments have
furnished many other details about the intensity and het-
erogeneity of the karst phenomenon, the velocity of karst
groundwater flow, and the relationship and overlap of iso-
tope data with hydrochemical ones.
In this study, particular attention was paid to the hydrol-
ogy and the successive catastrophic decreases of the Prepa
Lake levels that were observed during the period 1951–2020,
which—at its maximum—reached 9.58m above current lev-
els. The results of many investigations or papers were chal-
lenged, and different explanations or hypotheses provided in
an attempt to explain the lake level decrease can be summa-
rised as follows: (a) an increase of the transmissibility of the
karst aquifer that separates Lake Prespa from Lake Ohrid,
caused by geologic–tectonic reasons, resulting in the inten-
sification of the drainage of the first lake into the second
one; (b) increased, continuous and uncontrolled use of lake
water by the local population for agricultural, industrial and
other uses, and (c) change of climate conditions in recent
time. In conclusion, the authors believe that the problem
is very complex, and that the existing data are weak and
spread unevenly over the investigated area, both horizontally
and vertically. However, the climate changes that occurred
in the last 50years, as well as the non-effective manage-
ment of the water resources should be taken into account as
the main explanation of the registered water level decrease.
Also, it is important to establish at which rate the two most
relevant factors, i.e., the effects of climate change and the
uncontrolled management of water resources in all three
countries are likely to have influenced the rate of decline of
lake water levels.
Currently, there is no common legal framework and no
common criteria for assessing the causes of the depletion
of Lake Prespa. It is therefore recommended that a joint
consultative body is established to develop relevant man-
agement measures, with the contribution of highly qualified
local experts. The aim of such a body would be to:
I. Enhance cooperation in the management of the trans-
boundary water resources; and.
II. Ensure the sustainable use of water and other natural
resources associated with karst features in the region,
that also considered the likely influence of future cli-
mate changes on these resources.
The first step in the creation of such a joint consultative
body would be to collect and harmonise a large amount of
data and information relevant for the assessment and man-
agement of water resources in the catchment area of both
lakes, Ohrid and Prespa. The information that was gathered
in this study was not always complete and, in some cases,
included significant information gaps. To overcome this, a
new and common international water monitoring network
should be established and the information should be shared
between the three countries that have borders in the catch-
ment. Along with all the monitoring stations (hydrology,
climate elements) that are currently operational in the three
countries in question, new piezometers should be installed
along with a system for determining the overall rate of water
withdrawal that is pumped from the lakes, springs and bat-
teries of wells. Therefore, an urgent message of this study is
a request for the improvement of the groundwater monitor-
ing network throughout the region, and the need to intensify
capacity building in the public sector.
Some of the most important duties of the proposed joint
consultative body would be the following: sharing monitor-
ing data and enabling expert analysis on validated moni-
toring data; proposing measures to improve the water and
environmental situation in the catchment; advising changes
in legislation in order to harmonise some of the by-laws;
reporting to institutions in the countries and abroad on the
state of water resources in the catchment, and cooperation;
disseminating experience and lessons learned at various
educational levels; and technical capacity building and rais-
ing the awareness of the local population of the importance
of water and the dependent ecosystems’ protection from
pollution.
Declarations
Conflict of interest Not applicable to this manuscript.
Environmental Earth Sciences (2021) 80:708
1 3
708 Page 18 of 19
References
Aliaj Sh (2012) Neotectonics of Albanian (in Albanian). Tirana,
KLEAN, p 292
Aliaj Sh (1999) Neotectonics and seismotectonics of Korcha graben
region. In: Proceedings of conf. “Expert assessment of land sub-
sidence related to hydrogeological and engineering geological
conditions in regions of Sofia, Skopje and Tirana” Fourth working
group meeting, Sofia, UNESCO and BAS; pp 54–66
Amataj S, Anovski T, Eftimi R, Benishke R, Eftimi R, Gourcy L, Kola
L, Leontiadis I, Micevski E, StamosZoto A (2007) Tracer meth-
ods used to verify the hypothesis of Cvijic about the groundwa-
ter connection between Prespa and Ohrid lakes. Environ Geol
51:749–753
Andrianopulous A, Isajlovski N, Beska P (1997) Horizontal transport
of sediment materials. In: International Symposium “Towards
Integrated Conservation and Sustainable Development of Trans-
boundary Macro and Micro Prespa Lakes”, Korça, Albania, 24–26
October 1997; pp 65–67
Arsovski M (1997) Tectonics of Macedonia (in Macedonian). Pub-
lished by VEDA, Skopje
Anovski T, Andonovski B, Minceva B (1991) Study of the hydrologic
relationship between Ohrid and Prespa lakes. In: Proceedings
of IAEA International Symposium, 11–15 March1991, IAEA-
SM-319/ 62 p
Anovski T, Jovanovski N, Arsov LJ (1997) Rate determination of
water leakage from Prespa Lake. In: International Symposium
“TowardsIntegrated Conservation and Sustainable Development
of Transboundary Macro and Micro Prespa Lakes”, Korça - Alba-
nia, 24-26 October 1997; pp 29-31
Bakalowicz M (2005) Karst groundwater: a challenge for new
resources. Hydrogeol J 13:148–160
Benishke R, Goldscheider N, Smart C (2007) Tracer techniques. In:
Goldscheider N, Drew D (eds) Methods in karst hydrogeology.
Taylor and Francis, London, pp 147–170
Bonacci O (1987) Karst hydrology with special references to Dinaric
karst. Springer, Berlin
Bradlay M, Brown RM, Gonfiantini R, Payne BR, Prezewlocki K,
Sauzay G, Yen CK, Yurtsever Y (1972) Nuclear techniques in
groundwater hydrology. In: Groundwater Studies, Cap. 10. Pub-
lished by UNESCO 7, place de Fontenoy, 75700 Paris. Printed
by Imprimerie de la Manutention, Myenne. ISBN92-3-102206-7.
UNESCO, Paris
Chavkalovski I (1997) Hydrology of Prespa Lake. International
Symposium “Towards Integrated Conservation and Sustainable
Development of Transboundary Macro and Micro Prespa Lakes”,
Korça, Albania, 24–26 October 1997, pp 37–39
Clark ID, Fritz P (1997) Environmental isotopes in hydrogeology. CRC
Press/Lewis Publishers, Boca Raton
Coxon C (1999) Agriculture induced impacts. In: Drew D, Hötzl H
(eds) Karst hydrogeology and human activities. A.A Balkema,
Rotterdam, pp 37–80
Cvijić J (1906) Fundamental of geography and geology of Macedonia
and Serbia (in Serbian), Serbian Royal Academy, Spec. Ed. Books
1, 2, VIII+688, Belgrade
Drew D, Hötzl H (1999) Introduction. In: Drew D, Hötzl H (eds) Karst
hydrogeology and human activities. International contribution to
hydrogeology, vol 23. Balkema, Rotterdam, pp 3–12
Eftimi R (2010) Hydrogeological characteristics of Albania. AQUA-
mundi Am01012:079–092
Eftimi R (2020) Karst and karst water resources of Albania and their
management. Carbonates Evaporites 35:1–14
Eftimi R, Zoto J (1997). Isotope study of the connection of Ohrid and
Prespa lakes, International Symposium “Towards Integrated Con-
servation and Sustainable Development of Transboundary Macro
and Micro Prespa Lakes”, Korça, Albania, 24–26 October 1997;
pp. 32–37
Eftimi R, Zojer H (2015) Human impact on karst aquifers of Albania.
Environ Earth Sci 7:57–704
Eftimi R, Malik P (2019) Assessment of regional flow type and ground-
water sensitivity to pollution using hydrograph analyses and
hydrochemical data of the Selita and Blue Eye springs Albania.
Hydrogeol J. https://'doi.'org/'10.'1007/'s10040-'019-'01974-5
Fetter CW (1993) Contaminant hydrogeology. Macmillan, New York,
p 458
Ford D, Williams P (2007) Karst hydrogeology and geomorphology.
Wiley, Chichester, p 562
Gat JR, Dansgaard W (1970) Stable isotope survey of the freshwater
occurrences in Israel and Jordan Tift Valley. J Hydrol 16:177–212
Goldscheider N, Drew D (2007) Methods in karst hydrogeology. In:
Goldscheider N, Drew D (eds) International contribution to hydro-
geology, vol 26. Taylor and Francis, London, p 264
Guzelkovski D, Kotevski G (1977) Hydrogeological Map of Social-
ist Republic of Macedonia, scale 1:200 000, In: Rudarsko- Geo-
loshko Inzhinierstvo vo Rudarskite Grazhdenite i Vodostopanski
Objekti, 18–19 Oktomvri 2001, Struga
Hartmann A, Goldscheider N, Wagner T, Lange J, Weiler M (2014)
Karst water resources in a changing world: review of hydrologi-
cal modeling approaches. Rev Geophys. https://'doi.'org/'10.'1002/'
2013R'G0004'43
Hollis GE, Stevenson A (1997) The physical basis of the lake Mikri
Prespa systems: geology, climate, hydrology and water quality.
Hydrobiologia 351:1–19
IAEA (2003) Regional project REP/8/008, (2003) Study of Prespa
Lake using nuclear and related techniques. Final Report, Vienna
IAHS-IHLS (2004) Learning, teaching and applying techniques in
hydrology CD-ROM IAEA Vienna Isotope hydrology section.
IAEA, Vienna
Jordanovska B, Kunz MJ, Stafilov T, Wüest A (2010) Temporal vari-
ability in physic-chemical properties of St. Naum karts springs
feeding lake. Ohrid Ekol Zašt Život Sred 13(1–2):3–11
Jordanovska B, Stafilov T, Wüest A (2012) Assessment on physical-
chemical composition of surface karst springs feeding lake Ohrid.
Maced J Ecol Environ 16(12):19–25
Kanari K, Bregu V, Myftiu V, Tytymç P, Kelloli G & Kanina B (1997)
Problems of Micro Prespa Lake caused by Devoll river. In: Inter-
national Symposium “Towards Integrated Conservation and Sus-
tainable Development of Transboundary Macro and Micro Prespa
Lakes”, Korça, Albania, 24–26 October 1997; pp 181–182
Klinčarov S (1997) Geological-hydrogeological characteristics of
Prespa basin and their influence on the hydrogeological conditions
of the lake. In: International Symposium “Towards Integrated
Conser vation and Sustainable Development of Transboundary
Macro and Micro Prespa Lakes”, Korça, Albania, 24–26 October
1997; pp 79–83
Kresic N, Stevanovic Z (eds) (2010) Groundwater hydrology of
springs. Elsevier, Amsterdam, p 573
Lalovska F, Panov I (1997) Balance of the available waters in Prespa
Lake. In: Gjiknuri L, Miho A (eds) International Symposium
“Towards Integrated Conservation and Sustainable Development
of Transboundary Macro and Micro Prespa Lakes”, Korça, Alba-
nia, 24–26 October 1997; pp. 45–48
LaMoreaux PE, Wilson BM, Memon BA (1987) Guide to the hydrol-
ogy of carbonate rocks. In: LaMoreaux PE, Wilson BM, Memon
BA (eds) IHP Studies and reports in hydrology, vol 41. UNESCO,
Paris, p 345
Lauber U, Goldscheider N (2014) Use of artificial and natural tracers
to assess groundwater transit-time distribution and flow systems
in a high alpine karst system (Watterstein mountains, Germany).
Hydrogeol J 22:1807–1824
Environmental Earth Sciences (2021) 80:708
1 3
Page 19 of 19 708
Leng J, Wagner B, Bohem A, Panagiotopoulos K, Vane CH, Snel-
ling A, Haidon C, Woodley E, Vogel H, Zancheta G, Baneschi
I (2013) Understanding past climatic and hydrological variabil-
ity in the Mediterranean from Lake Prespa sediment isotope and
geochemical record over the Last Glacial cycle. Quatern Sci Rev
66:123–136
Leontiadis IL, Smyrniotis Ch (1986) Isotope hydrology study of the
Louros River-plains of Epirus, Greece. 5th International Sym-
posium On Groundwater Tracing, I.G.M.E., Athens; pp 75–90
Leontiadis IL, Smyrniotis Ch, Nikolaou E, Georgidis P (1997) Isotope
hydrology study of the major areas of Paramythia and Koroni,
Epirus, Greece. In: Proceedings of 5th international symposium
and field seminar on karst, waters and environmental impacts—
Antalya, Turkey, 10–20 September 1995, Balkema; pp 239–247
Löffler H, Schiller E, Kusel E, Krail H (1998) Lake Prespa, a European
natural monument, endangered by irrigation and eutrophication.
Hydrobiologia 384:69–74
Maloszewski PR, Benischke T, Harum H, Zojer H (1998) Estimation
of solute transport parameters in a karstic aquifer using artificial
tracer experiments. Int Contrib Hydrogeol 18:177–190
Matzinger A (2006) Is anthropogenic nutrient input jeopardizing
unique Lake Ohrid?—Mass flux analysis and management con-
sequences. Doct. Sciences Thesis. p 130
Matzinger A, Spirovski Z, Patceva Wüest S (2006) Sensitivity of
ancient Lake Ohrid to local anthropogenic impacts and global
warming. J Great Lakes Res 32:158–179
Meço S, Aliaj SH (2000) Geology of Albania. Gebrüder Borntraeger,
Stuttgart, p 246
Micevski E (2001) Geologic and hydrogeological characteristics of
Ohrid-Prespa region (in Macedonian). In: Rudarsko-Geoloshko
Inzhinierstvo vo Rudarskite Grazhdenite i Vodostopanski Objekti,
Struga, 18–19 Oktomvri 2001, pp 113–119
Milanović P (1981) Karst hydrogeology. Water Resources Publica-
tions, Littleton
Milanović P (2004) Water resources engineering in karst. CRC Press,
Boca Raton
Milevski J, Ristevski P and Tanusevska D (1997) In: International
Symposium “Towards Integrated Conservation and Sustainable
Development of Transboundary Macro and Micro Prespa Lakes”,
Korça, Albania, 24–26 October 1997; pp. 52–58
Nikolaou E, Pavlidis S, Katsanou K (2012) Aquifer systems of Epi-
rus, Greece: AN overview Kalavrita. Adv Res Aquat Environ
2011:425–433
Pano N (ed) (1984) Hydrology of Albania (in Albanian). Institute of
Hydrometeorology, Tirana, p 441
Parise M, Closson D, Gutiérrez F (2015) Stevanović Z (2015) Antici-
pating and managing engineering problems in the complex karst
environment. Environ Earth Sci 74:7823–7835
Payne BR, Leontiadis I, Dimitrulas Ch, Dounas A, Kallergis G, Morfis
A (1978) A study of the Kalamos Spring in Greece with environ-
mental isotopes. Water Resour Res 14(4):653–658
Popov V, Anovski T, Gospavić (2007) Sustainable management of
Prespa Lake. WIT Trans Eng Sci 58:71–79. https://'doi.'org/'10.'
2495/'EN070'081
Popov V, Anovska E, Arsov M, Amataj S, Kolaneci M, Stamos A,
Arsov L, Anovski T (2009) Study of Prespa-Ohrid lake system
using tracer experiments and the lake’s water balance. WIT Trans
Eng Sci 125:75–84
Popovska CO, Bonacci O (2007) Basic data on hydrology of Lakes
Ohrid and Prespa. Hydrol Process 21(5):658–664
Roberts N, Johnes MD, Benkaddour A, Wj E, Filippi MI, Frogley ML,
Lamb HF, Mj L, Jm R, Stain M, Stevens L, Valero- Garcés B,
Zanchetta G (2008) Stable isotope records of Late Quaternary
climate and hydrology from Mediterranean lakes: the ISOMED
synthethis. Quatern Sci Rev 27:2426–2441
Sappa G, Barbieri M, Ergul S, Ferrant F (2012) Hydrogeological con-
ceptual model of groundwater from carbonate aquifers using envi-
ronmental isotopes (18O, 2H) and chemical tracers: a case study in
Sothern Latium Region, Central Italy. J Water Res Prot 4:695–716
Selenica and Kolaneci (1997) The catastrophic falling of the water
level in Prespa lake. In: Gjiknuri L, Miho A (eds) International
Symposium “Towards Integrated Conservation and Sustainable
Development of Transboundary Macro and Micro Prespa Lakes”,
Korcha, Albania, 24–26 October 1997; pp 49–51
Shuster E, White W (1971) Seasonal fluctuations in the chemistry of
limestone springs: a possible mean for characterizing carbonate
aquifers. J Hydrol. https://'doi.'org/'10.'1016/'0022-'1694(71)'90001-1
Sibinović M (1987) Ezera Prespansko i Ohridsko. The SRC & Agency
of Water Resources of Macedonia, Skopje
Stamos A, Batsi A, Xantopoulou A (2014) Karstic aquifer systems and
relations of hydraulic communication with the Prespa Lakes. In:
The Tri-national Prespa Basin, Kalavrita, pp 527–534
Stevanović Z (2015) Karst aquifers: characterization and engineer-
ing. In: Stevanović Z (ed) Professional practice in earth sciences
series. Springer, Heidelberg, p 687
Stevanović Z, Filipović B (1994) Hydrogeology of carbonate rocks of
Carpathian-Balkanian. In: Stevanovic Z, Filipovic B (eds) Ground
waters in carbonate rocks of the Carpathian-Balkan Mountain
range. Spec.Ed. Of CBGA, Allston, pp 203–237
Stojov V (2011) Климатски и антропогени влијанија врз
водните резерви на тектонските езера [Climatic and
anthropogenic impacts on water reserves on the tectonic lakes].
MS Thesis, University of Ss. Cyr il and Methodius, Faculty of
Civil Engineering, Skopje, Macedonia. https://'www.'moepp.'gov.'
mk/'wp-'conte'nt/'uploa'ds/'2020/'12/'prespa_'klima_'antro'pogeni_'vlija'
nija_'20201'129.'pdf. Accessed 23 Jan 2021
Stojov V (2020) Climate and anthropogenic impacts on t he water
reserves of the Prespa Lake (Климaтcки и aнтpoпoгeни
влиjaниja вpз вoднитe peзepви нa Пpecпaнcкoтo Eзepo).
https://'www.'moepp.'gov.'mk/'wp-'conte'nt/'uploa'ds/'2020/'12/'prespa_'
klima_'antro'pogeni_'vlija'nija_'20201'129.'pdf. Accessed 23 Jan
2021
Stojov T, Stojov V, Trajanovska L (2004) Hydrometeorological analy-
ses typical for defining of water balance of Prespa Lake. BAL-
WOIS, Ohrid
Torbarov K (1976) Estimation of permeability and effective porosity
in karst on the basis of recession curve analysis. In: Hidrologija
i vodno bogatstvo krša, Zavod za hidrotehniku, Sarajevo, 97 p
Turc L (1954) Le bilan d’eau et des sols; relation entre les précipi-
tations, l’évaporation et l’écoulement. Troisièmes Journées de
l’hydraulique, Alger, pp 36–43
Wag ner B, Wilke T (2 011 ) E volu tio nar y and geo log ica l hi sto ry of B al-
kan lakes Ohrid and Prespa. Biogeosciences 8:995–998. https://'
www.'bioge'oscie'nces.'net/'8995/'2011/'doi:'10.'5194/'bg-8-'995-'2011.
Accessed 18 Jan 2021
White WB (1988) Geomorphology and hydrology of karst terrains.
Oxford University Press, New York
Xhomo A, Kodra A, Xhafa Z, Shallo M (2002) Geological Map of
Albania scale 1:200000. AGS, Tirana
Publisher's Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
A preview of this full-text is provided by Springer Nature.
Content available from Environmental Earth Sciences
This content is subject to copyright. Terms and conditions apply.
